Amazon Relational Database Service

RDS

You need a fully managed relational SQL database that handles replication, backups, and minor engine updates.

Regional serviceLearn it: Managed relational databasesConcept reference · no service console simulation

Reach for it when

Hosting transactional financial systems that require strict ACID compliance and complex multi-table joins.
Migrating standard MySQL or PostgreSQL databases to the cloud with minimal code modification.
Deploying applications that require foreign keys, complex views, and stored procedures.

Building highly ephemeral serverless workloads where database connections scale to thousands in seconds — use DynamoDB instead.
Storing massive volumes of unstructured log data or time-series events — use OpenSearch or DynamoDB instead.
Creating a globally distributed database that needs active-active writes in multiple regions — use Aurora Global Database instead.

Service	Pick it instead when
Aurora	Choose it when you need automated storage auto-scaling, faster replication, and serverless compute scaling.
DynamoDB	Choose it when you need horizontally scalable non-relational storage with no connection limits.

The model: Pay per hour of database instance operation based on instance class, plus storage and backup space.
The line item that surprises people: Allocating high provisioned IOPS storage (io1/io2) will bill a flat monthly fee even if the database is idle.

Exhausting database connections is common; Lambdas invoking RDS without a proxy like RDS Proxy can exhaust database connections; reuse connections and consider a proxy appropriate to the workload.
Modifying storage size or type can trigger a multi-hour optimization phase during which further modifications are blocked.
Automated backups require a backup window; during this window, I/O can be briefly suspended on single-AZ instances.

This page is a concept reference. Cost models are qualitative; confirm the current offering, Region and pricing before deploying.